Any recommendations on vendor / brand? I am experimenting with a couple of SR9s. The AP side has a 13db vpol 120deg panel. The client side (a mobile unit in my truck) has a 13db yagi and I get lousy reception even with LOS. 3 miles is the max and at -85 at that. I replaced the cable and checked my connections but no deal. I know that the performance has to be better. There is no noise (rural area). Perhaps is a crappy antenna. Dont know. Any opinions?

Something is wrong. I got -85ish with a 13dBi yagi on both ends, the one in my car facing the 90* wrong way balanced on my passenger seat and passenger mirror 1 mile away.

Right now I have no LOS, 13dBi yagi, lots of trees, 3.25 miles away, -86 to -90 signal. Stable until the cell phones start going.

Checked your antenna port? Noise in the area?

Very little noise. I sat running a scan last night at the AP side and picked up nothing in about 30 minutes. I then did a "channel utilization scan" and the noise would move from -105 to a high of -85 for brief periods with the more common reading been in the -96 region. I upgraded to v3.rc last night on the AP and will update my truck today and see if I can narrow down the problem. I really was expecting 5 mile NLOS performance in the -80 range.

As I mentioned above, my problems come from a cell tower nearby. Once they start getting a lot of traffic, my noise floor drops to -85ish as well, then signal gets really bad.

Keep in mind, at least in the US, licensed frequencies are putting out a lot more power in tighter frequency ranges, just an example, 4w on 2KHz channels, and those signals carry and do cause interference in this band.

I was just about ready to buy a cavity filter, but my test link time is up.

How high is your AP? The higher it is, the more interference you'll see, the more likely you'll need the cavity filter. You need to determine what frequency the noise is coming from. FCC site helped me in that regard.

Hmm, it must be understood that with the 900mhz band the read-outs from the mikrotik's noise floor figure are essentially and somewhat useless. (You would need a spec. any. to find out what the REAL noise floor is) However if your area is in fact clean (doubted) then you would wanna check that the polarizations of the antennas are identical (this can easily take away 25+db). The theoretical signal you would acheive with a 900mhz 5 mile link would be (-)~69 assuming that you transmitting at half power (14dbm; full power would yeild (-)~55) Also you might wanna decrease the size of the channel to perhaps 1/4 width as it usually helps a great deal with interference would give you 50% (3db increase) more power.


Regards

Yea, noise floor is relative, and out of band noise(as well as non 802.11 noise I assume) isn't recognized.

Im currently using 5mhz bandwidth. My tower sits on a hilltop about 100 feet above the surrounding valleys (customers) it is behind a small tree obstruction (about 10 feet across). I am going to try some downtilt and see what happens.

The 900 mhz solution is somewhat picky in when and where it works. I have done 11 mile NLOS shots with a 13 dbi yagi and a 14 dbi sector with signal in the mid to high -80's. I have also done site surveys, 1/4 mile from the tower and got nothing more than -95.

Who is the cellular carrier in your area? Nextel is the only one that gets close to you on the frequency. Most of the other B-Band carriers stop somewhere below 879 mhz. They have bandpass filters built into the antennas and cavity filters in the CDMA or GSM bay stations. The likely hood of them interfering with you is pretty unlikely. Now, if the cellular carrier has taken some lightning hits or some other surge that may have compromised the equipment, then maybe it could affect you. Look for SCADA radios on pipelines and power company sub stations. These are also in the unlicensed band and are more likely causes of interference.

We have been experimenting with 900 using the SR9 card. Without a cavity we had a noise floor that jumped from the low -80's to somewhere in the -90's depending on exact frequency we used and time of day. Busy cell sites and several 900 meg paging sites surround this particular area. We installed a Ubiquity bandpass filter because it has very steep sides and narrow bandpass. Many others we looked at had too wide a "window".

The results we have seen is that the noise floor has become more stable and stays between -92dB and -97dB. I just checked it at 9pm and it is at -95. This has made a much more dependable system. Our AP is on top of a grain elevator and well above the tree line, but the customers are well below tree line in many cases. We can reach customers now that were not reachable before.

As has been mentioned many times, 900 is tricky and will not work in every instance, but the use of a narrow filter can be beneficial. I hope this info will help someone.

Check FCC's site for licensed systems in your area.
One cell phone company here uses a band around 849 and a second band that tops out at 892. That is the one causing my interference, but that was a temp link.

Also, the power company here uses 930MHz for their monitoring system at the substations. I've never seen a blip at 930, but lots of noise just below 900MHz.

Where would I look in the FCC site? I did a search but come up with thousands of documents.

http://wireless.fcc.gov/uls/

Use the license search, Geosearch.

Takes a bit to find what you are looking for as often the licenses are by location of the licensee, not the transmitter. Geosearch, if memory serves, goes by the location of the transmitter in x miles/km from a certain location.